1. Field of the Invention
The present invention relates to optical disk apparatuses reproducing signals from optical disks, and more particularly to an optical disk apparatus which reproduces signals from an optical disk by accurately converting sampling data of the reproduction signal from multilevel data to binary data even if the reproduction signal varies in amplitude.
2. Description of the Background Art
Recently, with an increase in recording density of digital disks such as magneto-optical disks, optical disks and phase change disks, a partial response (PR) coding method has been widely used which utilizes intersymbol interference by matching transmission signals with transmission path characteristics. Viterbi decoding (maximum likelihood decoding) is known as a decoding method of the PR coding.
In the Viterbi decoding for a conventional optical disk apparatus, a radio frequency (RF) signal being a reproduction signal obtained from an optical disk is sampled, and an error or square error (branch metric) between the sampled digital data and an expected value is calculated. From this square error, an accumulative error (path metric) is calculated for each data stream, and the data stream whose accumulative error is the smallest is selected as the most likely data stream. The RF signal converted to digital data is then converted from multilevel data to binary data.
In the conventional optical disk apparatus, however, in the case where the expected value for use in the Viterbi decoding is fixed to a constant value, the branch metric or the path metric, which is calculated based on the square error between the amplitude of the RF signal and the expected value of the amplitude, would not always allow precise Viterbi decoding, due to variation in amplitude of the RF signal attributable to unevenness in rotation speed, eccentricity or tilt of the digital disk.
Japanese Patent Laying-Open No. 9-320206 discloses a method for performing accurate Viterbi decoding despite such variation in amplitude of the RF signal, wherein a reproduction signal obtained from an optical disk is sampled and, from among the sampled data, a zero-cross sample value located nearest the zero-cross point and values preceding and succeeding the zero-cross sample value are employed as expected values in a Viterbi decoder.
According to the method described in the Japanese Patent Laying-Open No. 9-320206, however, the sampling values in the region where the reproduction signal obtained from the optical disk exhibits largest variation in amplitude have been used as the expected values. This hinders stable setting of the expected values, and as a result, Viterbi decoding cannot be performed with high accuracy.